Dynamic Bayesian models as an alternative to the estimation of operational risk measures

Renato da Silva Carvalho, Hélio S. Migon, Marina Silva Paez

Abstract

ABSTRACT

In this paper our aim is to introduce a new class of models, under a Bayesian point of view, as an alternative to the treatment of operational risk assessment issues. We show that some drawbacks in traditional modeling is overcome. During the past few years, the interest in and need for useful, accurate and robust measures of operational risk have grown. The most widely accepted measure is the operational value-at-risk. In this paper, alternative models are compared for the estimation of operational risk measures in a historical series of administrative fines imposed by ANATEL (Brazilian Telecommunications Agency) to a specific company. One of the advantages of the proposed methodology is that it is able to incorporate a dynamic component in time, which helps to explain the variability of the series.